Using artificial neural network models to assess water quality in water distribution networks

Abstract

The purpose of the research is to assess chlorine concentration in WDS using statistical models based on ANN in combination with Monte-Carlo. This approach offers advantages in contrast to the generally use methods for modeling of chlorine decay in drinking water systems until now. The model was tested on one specific location using the hydraulic and water quality parameters such as flow, pH, temperature, etc. The model allows forecasting chlorine concentration at selected nodes of the water supply system. The results obtained in these selected nodes allow then to compare the chlorine concentration with EPANET in the system under assessment.

The purpose of the research is to assess chlorine concentration in WDS using statistical models based on ANN in combination with Monte-Carlo. This approach offers advantages in contrast to the generally use methods for modeling of chlorine decay in drinking water systems until now. The model was tested on one specific location using the hydraulic and water quality parameters such as flow, pH, temperature, etc. The model allows forecasting chlorine concentration at selected nodes of the water supply system. The results obtained in these selected nodes allow then to compare the chlorine concentration with EPANET in the system under assessment.